In a magnetic disk device, when additional data are written to the device and an error in reading information related to data management in the device occurs, due to the error, the amount of data corrupted is increased. An error detection and correction technique is devised and applied as a countermeasure against the error. Particularly, one means for error recovery is a device configured to add check value data, e.g., XOR data based on XOR (exclusive logical sum) calculated parity data.
In the related art, in the error recovering method which uses XOR data, all sectors allocated along a certain track are set as one user block and the XOR data is registered in the last check value sector (XOR sector). Accordingly, each time one or more sectors on a track are rewritten, the XOR sector for that track is also rewritten. Generally, during read and write processing of consecutive blocks, a phase of a sector position is shifted on each track circumferentially relative to adjacent tracks so as to optimize timing when reading multiple contiguous blocks of data. Therefore, since sector numbers are not radially aligned with each other, the rewriting of the XOR sectors has a strong effect, via adjacent track interference (ATI), on adjacent sectors in the radial direction. This causes an increase in the number of times data recovery processing (e.g., refresh processing) needs to be performed on the track that includes the adjacent sectors.